A Survey on Mining of Weakly Labeled Web Facial Images and Annotation

Now-a-days, lots of images are generated and uploaded on the internet due to the popularity of digital cameras and social media tools. Some of these facial images are tagged correctly, but most of them are not tagged correctly, so facial annotation came into the picture. There are many applications found in online photo album management, real world management system and in multimedia to correctly identify the person in the video domain. On the World Wide Web (WWW), most images are wrongly labeled so it asserts towards a strong need of a system that strongly performs mining and annotation of these weakly labeled images. Face annotation associated with face detection and recognition recently researchers interests in mining weakly-labeled facial pictures on the net to resolve research challenges in computer vision and image understanding. This paper provides varied techniques or ways that are used for annotating facial images. One testing issue problem for search-based face annotation plan is the way to effectively perform tagging by retrieving the list of most comparative facial pictures and their weak labels that are frequently noisy and deficient. The main objective of the proposed system is to assign right name labels to a given query facial picture.